| Author |
Message |
baldgeezer
Guest
|
 Posted:
Thu Oct 20, 2005 8:51 am Post subject:
Suspect Onepoint database on cluster |
|
|
OK, not had this before. I've installed the database component on a SQL2000
SP4 cluster (active/passive). When the SQL virtual server is running on the
active node, the Onepoint database is fine. Fail over the cluster to the
passive node and the OnePoint database is marked as suspect. However - now
this is the really weird bit, running the stored procedure sp_resetstatus
against the OnePoint database when on the passive node returns the
following:
Prior to updating sysdatabases entry for database 'onepoint', mode = 0 and
status = 24 (status suspect_bit = 0).
No row in sysdatabases was updated because mode and status are already
correctly reset. No error and no changes made.
So - is it really suspect or what? Anyone seen this before and got any
suggestions for getting it working on the passive node properly?
-Dave |
|
| Back to top |
|
 |
baldgeezer
Guest
|
| Posted:
Thu Oct 20, 2005 8:51 am Post subject:
Re: Suspect Onepoint database on cluster |
|
|
And before anyone asks - yes I've restarted the SQL services following
failover. I've checked the SQL logs on the failover node and I get the
following:
2005-10-20 09:08:51.19 spid12 Starting up database 'ReportServer'.
2005-10-20 09:08:51.21 spid10 udopen: Operating system error 5(error not
found) during the creation/opening of physical device I:\Program
Files\Microsoft SQL Server\MSSQL\Data\EeaLog.ldf.
2005-10-20 09:08:51.21 spid10 FCB::Open failed: Could not open device
I:\Program Files\Microsoft SQL Server\MSSQL\Data\EeaLog.ldf for virtual
device number (VDN) 2.
2005-10-20 09:08:51.21 spid13 Starting up database 'ReportServerTempDB'.
2005-10-20 09:08:51.21 spid10 Device activation error. The physical file
name 'I:\Program Files\Microsoft SQL Server\MSSQL\Data\EeaLog.ldf' may be
incorrect.
2005-10-20 09:08:51.21 spid10 Device activation error. The physical file
name 'I:\Program Files\Microsoft SQL Server\MSSQL\Data\EeaLog.ldf' may be
incorrect.
So - as per the norm I've checked that the I: drive is accessible and it is
(it is in the same group as the J: drive (data) ) and restarted the SQL
services. Still no joy.
"baldgeezer" <baldgeezer@ntlworld.com> wrote in message
news:T4J5f.14992$pp5.2251@news.cpqcorp.net...
| Quote: | OK, not had this before. I've installed the database component on a
SQL2000 SP4 cluster (active/passive). When the SQL virtual server is
running on the active node, the Onepoint database is fine. Fail over the
cluster to the passive node and the OnePoint database is marked as
suspect. However - now this is the really weird bit, running the stored
procedure sp_resetstatus against the OnePoint database when on the passive
node returns the following:
Prior to updating sysdatabases entry for database 'onepoint', mode = 0 and
status = 24 (status suspect_bit = 0).
No row in sysdatabases was updated because mode and status are already
correctly reset. No error and no changes made.
So - is it really suspect or what? Anyone seen this before and got any
suggestions for getting it working on the passive node properly?
-Dave
|
|
|
| Back to top |
|
|
baldgeezer
Guest
|
| Posted:
Thu Oct 20, 2005 8:51 am Post subject:
Re: Suspect Onepoint database on cluster |
|
|
Ok - just done another test with another DB and I get the same behaviour so
it is not MOM but SQL. This DB previously had it's Log and Data files on the
same disk, moving the log to a different disk in the same cluster group
causes the issue. More digging required!!!!
"baldgeezer" <baldgeezer@ntlworld.com> wrote in message
news:mcJ5f.14993$Ha5.10333@news.cpqcorp.net...
| Quote: | And before anyone asks - yes I've restarted the SQL services following
failover. I've checked the SQL logs on the failover node and I get the
following:
2005-10-20 09:08:51.19 spid12 Starting up database 'ReportServer'.
2005-10-20 09:08:51.21 spid10 udopen: Operating system error 5(error
not found) during the creation/opening of physical device I:\Program
Files\Microsoft SQL Server\MSSQL\Data\EeaLog.ldf.
2005-10-20 09:08:51.21 spid10 FCB::Open failed: Could not open device
I:\Program Files\Microsoft SQL Server\MSSQL\Data\EeaLog.ldf for virtual
device number (VDN) 2.
2005-10-20 09:08:51.21 spid13 Starting up database
'ReportServerTempDB'.
2005-10-20 09:08:51.21 spid10 Device activation error. The physical
file name 'I:\Program Files\Microsoft SQL Server\MSSQL\Data\EeaLog.ldf'
may be incorrect.
2005-10-20 09:08:51.21 spid10 Device activation error. The physical
file name 'I:\Program Files\Microsoft SQL Server\MSSQL\Data\EeaLog.ldf'
may be incorrect.
So - as per the norm I've checked that the I: drive is accessible and it
is (it is in the same group as the J: drive (data) ) and restarted the SQL
services. Still no joy.
"baldgeezer" <baldgeezer@ntlworld.com> wrote in message
news:T4J5f.14992$pp5.2251@news.cpqcorp.net...
OK, not had this before. I've installed the database component on a
SQL2000 SP4 cluster (active/passive). When the SQL virtual server is
running on the active node, the Onepoint database is fine. Fail over the
cluster to the passive node and the OnePoint database is marked as
suspect. However - now this is the really weird bit, running the stored
procedure sp_resetstatus against the OnePoint database when on the
passive node returns the following:
Prior to updating sysdatabases entry for database 'onepoint', mode = 0
and status = 24 (status suspect_bit = 0).
No row in sysdatabases was updated because mode and status are already
correctly reset. No error and no changes made.
So - is it really suspect or what? Anyone seen this before and got any
suggestions for getting it working on the passive node properly?
-Dave
|
|
|
| Back to top |
|
|
Arie de Haan
Guest
|
| Posted:
Tue Oct 25, 2005 8:51 am Post subject:
Re: Suspect Onepoint database on cluster |
|
|
| Quote: | 2005-10-20 09:08:51.19 spid12 Starting up database 'ReportServer'.
2005-10-20 09:08:51.21 spid10 udopen: Operating system error 5(error
not found) during the creation/opening of physical device I:\Program
|
So it cannot access the file exclusively.. did you have the resources
dependecy correctly set? Does a virusscanner run and locks it at the
moment of failover? read only attribute set?
Anyways something is preventing SQL to exclusively lock the file, i
think your troubleshooting should go in that direction
http://support.microsoft.com/default.aspx?scid=kb;en-us;180500
--
Greetz,
Arie
This posting is provide "AS IS" with no guarantees, warranties, rigths
etc. |
|
| Back to top |
|
|
baldgeezer
Guest
|
| Posted:
Thu Nov 10, 2005 5:51 pm Post subject:
Re: Suspect Onepoint database on cluster |
|
|
Believe it or not - upgrading the cluster to SQL2005 solved the problem!
"baldgeezer" <baldgeezer@ntlworld.com> wrote in message
news:TjJ5f.14994$7k5.3938@news.cpqcorp.net...
| Quote: | Ok - just done another test with another DB and I get the same behaviour
so it is not MOM but SQL. This DB previously had it's Log and Data files
on the same disk, moving the log to a different disk in the same cluster
group causes the issue. More digging required!!!!
"baldgeezer" <baldgeezer@ntlworld.com> wrote in message
news:mcJ5f.14993$Ha5.10333@news.cpqcorp.net...
And before anyone asks - yes I've restarted the SQL services following
failover. I've checked the SQL logs on the failover node and I get the
following:
2005-10-20 09:08:51.19 spid12 Starting up database 'ReportServer'.
2005-10-20 09:08:51.21 spid10 udopen: Operating system error 5(error
not found) during the creation/opening of physical device I:\Program
Files\Microsoft SQL Server\MSSQL\Data\EeaLog.ldf.
2005-10-20 09:08:51.21 spid10 FCB::Open failed: Could not open device
I:\Program Files\Microsoft SQL Server\MSSQL\Data\EeaLog.ldf for virtual
device number (VDN) 2.
2005-10-20 09:08:51.21 spid13 Starting up database
'ReportServerTempDB'.
2005-10-20 09:08:51.21 spid10 Device activation error. The physical
file name 'I:\Program Files\Microsoft SQL Server\MSSQL\Data\EeaLog.ldf'
may be incorrect.
2005-10-20 09:08:51.21 spid10 Device activation error. The physical
file name 'I:\Program Files\Microsoft SQL Server\MSSQL\Data\EeaLog.ldf'
may be incorrect.
So - as per the norm I've checked that the I: drive is accessible and it
is (it is in the same group as the J: drive (data) ) and restarted the
SQL services. Still no joy.
"baldgeezer" <baldgeezer@ntlworld.com> wrote in message
news:T4J5f.14992$pp5.2251@news.cpqcorp.net...
OK, not had this before. I've installed the database component on a
SQL2000 SP4 cluster (active/passive). When the SQL virtual server is
running on the active node, the Onepoint database is fine. Fail over the
cluster to the passive node and the OnePoint database is marked as
suspect. However - now this is the really weird bit, running the stored
procedure sp_resetstatus against the OnePoint database when on the
passive node returns the following:
Prior to updating sysdatabases entry for database 'onepoint', mode = 0
and status = 24 (status suspect_bit = 0).
No row in sysdatabases was updated because mode and status are already
correctly reset. No error and no changes made.
So - is it really suspect or what? Anyone seen this before and got any
suggestions for getting it working on the passive node properly?
-Dave
|
|
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in